检索结果-内蒙古大学图书馆

ChemDFM-X: towards large multimodal model for chemistry

science China(Information sciences) 2024年第12期67卷 99-100页

作者： Zihan ZHAO Bo CHEN Jingpiao LI Lu CHEN Liyang WEN Pengyu WANG Zichen ZHU Danyang ZHANG Yansi LI Zhongyang Dai Xin CHEN Kai YU X-LANCE Lab Department of Computer Science and EngineeringMoE Key Lab of Artificial Intelligence AI Institute Shanghai Jiao Tong University Suzhou Laboratory

Chemistry, as a naturally multimodal discipline, plays a crucial role in various vital fields such as pharmaceutical research and material manufacturing. Therefore, research on artificial intelligence(ai) for chemistry has garnered increasing attention. Despite the rapid development, most of the chemical ai models today mainly focus on single tasks with unimodal input [1].

关键词：

来源：评论

学校读者我要写书评

暂无评论

ai-Driven Resource and Communication-Aware Virtual Machine Placement Using Multi-Objective Swarm Optimization for Enhanced Efficiency in Cloud-Based Smart Manufacturing

引用

computers, Materials & Continua 2024年第12期81卷 4743-4756页

作者： Praveena Nuthakki Pavan Kumar T. Musaed Alhussein Muhammad Shahid Anwar Khursheed Aurangzeb Leenendra Chowdary Gunnam Department of Computer Science&Engineering Koneru Lakshmaiah Education FoundationGuntur522302India Department of Computer Engineering College of Computer and Information SciencesKing Saud UniversityRiyadh11543Saudi Arabia Department of AI and Software Gachon UniversitySeongnam-Si13120Republic of Korea Department of Electronics and Communication Engineering SRM UniversityAmaravati522502India

Cloud computing has emerged as a vital platform for processing resource-intensive workloads in smart manu-facturing environments,enabling scalable and flexible access to remote data centers over the *** these environments,Virtual Machines(VMs)are employed to manage workloads,with their optimal placement on Physical Machines(PMs)being crucial for maximizing resource ***,achieving high resource utilization in cloud data centers remains a challenge due to multiple conflicting objectives,particularly in scenarios involving inter-VM communication dependencies,which are common in smart manufacturing *** manuscript presents an ai-driven approach utilizing a modified Multi-Objective Particle Swarm Optimization(MOPSO)algorithm,enhanced with improved mutation and crossover operators,to efficiently place *** approach aims to minimize the impact on networking devices during inter-VM communication while enhancing resource *** proposed algorithm is benchmarked against other multi-objective algorithms,such as Multi-Objective Evolutionary Algorithm with Decomposition(MOEA/D),demonstrating its superiority in optimizing resource allocation in cloud-based environments for smart manufacturing.

关键词： Resource utilization smart manufacturing efficiency inter VM communication virtual machine placement cloud computing multi-objective optimization

来源：评论

学校读者我要写书评

暂无评论

How Does Software Prefetching Work on GPU Query Processing? 20

How Does Software Prefetching Work on GPU Query Processing?

引用

20th International Workshop on Data Management on New Hardware, DaMoN 2024

作者： Deng, Yangshen Chen, Shiwen Hong, Zhaoyang Tang, Bo Department of Computer Science and Engineering Southern University of Science and Technology AlayaDB Ai Guangdong Shenzhen China

ISBN: (纸本)9798400706677

Improving the performance of GPU query processing is a well-studied problem in database community. However, its performance is still unsatisfactory due to the low utilization of GPU memory bandwidth. In the literature, employing software prefetching techniques to improve the bandwidth utilization is a common practice in CPU database as it overlaps computation cost and memory access latency. However, it was ignored by GPU database even though the software prefetching ability has been provided by modern GPU architecture (i.e., from NVIDIA Ampere). In order to investigate the effectiveness of software prefetching techniques on GPU query processing, we implement four software prefetching algorithms on GPU, i.e., Group Prefetch (GP), Software-Pipelined Prefetch (SPP), Asynchronous Memory Access Chaining (AMAC) and Interleaved Multi-Vectorizing (IMV) in the work. We then adapt them on hash join probe and BTree search tasks with a suite of optimizations. Last, we conduct comprehensive experiments and evaluate the performance of them. The results confirm the superiority of software prefetching techniques on GPU query processing. Specifically, they can achieve up to 1.19X speedup on hash join probe and 1.31X speedup on BTree search when compared with the implementations without software prefetching. © 2024 ACM.

关键词： Graphics processing unit

来源：评论

学校读者我要写书评

暂无评论

Proactive Purity: Predicting Water Quality with Precision

Proactive Purity: Predicting Water Quality with Precision

引用

Trends in Material science and Inventive Materials (ICTMIM), International Conference on

作者： K. Sandhya G. Ashitha B. Sharath Kumar B. Varun Kumar T. Bharadwaj Department of Computer Science and Engineering (AI & ML) Vignan Institute of Technology and Science Nalgonda Telangana

ISBN: (数字)9798331501488

ISBN: (纸本)9798331501495

Water pollution, driven by industrialization and urbanization, threatens public health, aquatic ecosystems, and livelihoods worldwide. Rivers, being the primary source of water for drinking, agriculture, and industry, are highly vulnerable to contamination from untreated sewage, industrial discharge, and agricultural runoff. Traditional water quality monitoring methods are expensive, time-consuming, and lack real-time *** study proposes a technology-driven water quality monitoring system capable of processing large data volumes in real-time, identifying pollution patterns, and supporting timely interventions. By integrating advanced monitoring techniques, the system enhances decision-making for water management, ensuring sustainability and ecological balance. The research highlights the critical role of technology in addressing global water pollution balance, further underlining the very important linkage between technology and *** data volumes from source water can be processed by this system, identifying patterns and trends that indicate fluctuation in water quality. this system is particularly beneficial in areas that experience seasonal and man-made causes of change in water quality. Besides its technological achievements, this project is a giant leap toward the management of water *** technology provides authority to decision makers. who together with environmentalists and enables easy accessibility for various tasks to make sensible decisions for drinking water supply as well as irrigation systems and pollution prevention measures for water sources. measures. It has great importance as it combines science and technology in saving vital natural resources and maintaining long-term ecological balance by tackling one of the most urgent worldwide environmental issues.

关键词： Machine learning algorithms Water quality Prediction algorithms Market research Water pollution Real-time systems Pollution measurement Public healthcare Water resources Monitoring

来源：评论

学校读者我要写书评

暂无评论

Detection and Prevention of Workflow Attrition

Detection and Prevention of Workflow Attrition

引用

Pervasive Computing and Social Networking (ICPCSN), International Conference on

作者： J. R. V. Jeny V. Sunitha Chowdhary A. Nischay Reddy P. Nikhil Goud Department of Computer Science and Engineering (AI & ML) Vignan Institute of Technology and Science Nalgonda Telangana

ISBN: (数字)9798331535193

ISBN: (纸本)9798331535209

Each business acknowledges the importance of employees in building and sustaining competitive advantage. The turnover of employees should, therefore, be seen as something that negatively affects business growth. It is in this process that sound decisions are made regarding management. Its role may very well be of utmost importance in any planning prepared. Worker turnover has been a known factor, and decisions in management call for sound judgment to retain better employees. Today's firms have a significant interest in finding the causes and circumstances behind employee turnover to enable its reduction. Some causes contribute to instability. In this manner, predicting the likelihood of employee turnover and primarily identifying causes of turnover are vital organizational objectives to optimize HR techniques. Interests, artificial intelligence (ai), machine learning, and deep learning have been actively applied to anticipate instability likelihood using automated techniques. This paper aims to utilize machine learning and deep learning models and compare them to attain the highest possible accuracy. We intend to achieve an accuracy of 94%, which is higher than the previous best of 92%, using the ANN and CNN algorithms.

关键词： Deep learning Accuracy Social networking (online) Computational modeling Prevention and mitigation Stacking Predictive models Prediction algorithms Robustness Planning

来源：评论

学校读者我要写书评

暂无评论

PaySafe Al: Intelligent Fraud Detection for UPI Transactions using Machine Learning

PaySafe Al: Intelligent Fraud Detection for UPI Transactions...

引用

International Conference on ent Computing and Control Systems (ICICCS)

作者： Markala Anjali Maddela Ajay Mukkerla Saiteja Baddula Omkar Yadav Department of Computer Science and Engineering (AI & ML) Vignan Institute of Technology and Science Nalgonda Telangana

ISBN: (数字)9798331512088

ISBN: (纸本)9798331512095

In the financial world, it is impoartant to have a sound understand of systems against fraud, as they help detect activities such as anomalous and malicious transactions, through the analysis of activity engagement. Decision Tree are the recent form of deep learning models which, in more recent years, have shown advanced capability in the identification of fraudulent transactions, however, sensitivity and accuracy when implemented in real time systems remain an issue. In this research, we provide a new idea of fraud detection of UPI transactions that utilizes a Decision Tree based sequence modeling framework. The model focuses on normalization sets and typical cardholder activities to effectively learn how to detect anomalies in UPI transactions. It then analyzes the sequences of a person’s activities and adds on the probabilities, to then classify if the transaction is fraudulent or not. Several measures are established, however, in order to aid in minimizing false positives, because the aim is to ensure valid transactions are still processed. The experimental results support the proposed fraud detection system’s potential usefulness in monitoring actual UPI transactions and it also shows how it outperforms current systems in related studies. Our model facilitates the UPI transaction environment equally and addresses the fine balance in detection efficiency, accuracy and transaction speed by proposing the integration of deep learning techniques.

关键词： Support vector machines Deep learning Logistic regression Accuracy Sensitivity Predictive models Fraud Decision trees Security Standards

来源：评论

学校读者我要写书评

暂无评论

ai-Powered Ingredient Detector for Allergies: Enhancing Food Safety Through Natural Language Processing and computer Vision

AI-Powered Ingredient Detector for Allergies: Enhancing Food...

引用

International Conference on computer and Communications Technologies (ICCCT)

作者： Suwetha K Abhiram R Crispin R Nagaraj R Department of Computer Science and Engineering (AI & ML) Sri Sairam Engineering College Chennai India

ISBN: (数字)9798331537579

ISBN: (纸本)9798331537586

Food allergies pose a significant health risk, affecting millions worldwide and potentially leading to severe reactions, including anaphylaxis. Identifying allergens in food products is crucial for individuals with allergies. However, the increasing complexity of food labeling and hidden allergens makes this process challenging, time-consuming, and error-prone. This paper presents the ai-powered Ingredient Detector for Allergies, an innovative application designed to streamline allergen detection using Artificial Intelligence (ai). The system enables users to input a food item, scan a barcode, or upload an image for ingredient analysis. By leveraging Natural Language Processing (NLP) and computer vision, it accurately identifies potential allergens and provides real-time alerts. This application is integrated with a comprehensive allergen database, which ensures precise detection and enhances food safety. By offering a user-friendly, intelligent, and efficient solution, this ai-driven system empowers individuals with food allergies to make informed dietary choices confidently and efficiently.

关键词： computer vision Databases Scalability Detectors Streaming media Natural language processing Real-time systems Food safety Artificial intelligence Allergies

来源：评论

学校读者我要写书评

暂无评论

Invisible Shield: Real-Time Industrial Threats Detection for IoT-based Systems

Invisible Shield: Real-Time Industrial Threats Detection for...

引用

International Conference on ent Computing and Control Systems (ICICCS)

作者： Kadaganchi Sudhakar Pasula Manish Mavillapally Rohit Shaik Kaleem Ahmed Batharaju Sneha Department of Computer Science and Engineering (AI & ML) Vignan Institute of Technology and Science Hyderabad Telangana

ISBN: (数字)9798331512088

ISBN: (纸本)9798331512095

The Industrial Internet of Things' (IIoT) explosive growth has significantly changed industrial environments by linking smart devices via Supervisory Control and Data Acquisition (SCADA) systems. This integration also introduces significant cyber security vulnerabilities even though it brings about flexibility, resource efficiency, and operational agility. Current IDS using traditional machine learning fail to classify cyber attacks precisely because of complexity in data, limited availability, and mislabeling issues. To overcome these challenges, this study presents a scalable and effective ensemble detection framework that utilizes Pyramidal Recurrent Units (PRUs) and Decision Tree (DT) models. This framework aims to detect and counter cyber attacks across large IIoT networks, especially in SCADA-based environments. The propose of developing this project is to ensure scalable and efficient Deep Learning and Decision Tree based ensemble cyber attack detection framework to resolve trustworthiness issues in the SCADA based IIoT networks. Our proposed detection method can be applied to various IIoT domains. It is easy to implement and deploy, improving efficiency and accuracy while addressing the limitations of earlier *** framework enhances the security of SCADA-based IIoT systems, making industrial networks more reliable, trustworthy, and resilient. The results indicates that the system performs well across different network settings, showcasing its adaptability and robustness in detecting IIoT-based SCADA systems.

关键词： Deep learning Accuracy SCADA systems Feature extraction Threat assessment Complexity theory Security Decision trees Cyberattack Industrial Internet of Things

来源：评论

学校读者我要写书评

暂无评论

Enhancing Privacy in Real-Time Stream Processing: Federated Transfer Learning Approaches

Enhancing Privacy in Real-Time Stream Processing: Federated ...

引用

2024 IEEE International Workshop on Metrology for Living Environment, MetroLivEnv 2024

作者： Jog, Shwetha Palaniappan, Damodharan Jabbar, M.A. Marwadi University Department of Computer Engineering Rajkot India Marwadi University Department of Information Technology Rajkot India Vardhaman College of Engineering Department of CSE-AI&ML Hyderabad India

ISBN: (纸本)9798350385014

In the era of data ubiquity, safeguarding privacy while harnessing the power of real-time streaming presents a critical challenge. This study delves into the intersection of privacy preservation, real-time streaming, and federated transfer learning to devise robust solutions. First, we elucidate the essence of data privacy, real-time streaming, and federated transfer learning. Subsequently, we propose an innovative approach that seamlessly integrates privacy considerations into the fabric of real-time streaming analytics through federated learning models. Leveraging this framework, we ascertain the optimal model selection process tailored to the specific data characteristics and user requirements. Our analysis encompasses a comparative evaluation of five prominent models: Differential Privacy, Secure Multi-party Computation, Homomorphic Encryption, Federated Learning with User-level Differential Privacy, and Decentralized Federated Learning. Through empirical examination, we discern the most suitable model for preserving privacy in real-time streaming contexts, thereby advancing the discourse on privacy-preserving machine learning paradigms. © 2024 IEEE.

关键词： Privacy-preserving techniques

来源：评论

学校读者我要写书评

暂无评论

ControlVideo: conditional control for one-shot text-driven video editing and beyond

引用

science China(Information sciences) 2025年第3期68卷 150-162页

作者： Min ZHAO Rongzhen WANG Fan BAO Chongxuan LI Jun ZHU Department of Computer Science and Technology Institute for AI Tsinghua-Bosch Joint ML CenterTsinghua Laboratory of Brain and Intelligence Lab Tsinghua University ShengShu Technology Gaoling School of Artificial Intelligence Renmin University of China Beijing Key Laboratory of Big Data Management and Analysis Methods Pazhou Laboratory (Huangpu)

This paper presents ControlVideo for text-driven video editing — generating a video that aligns with a given text while preserving the structure of the source video. Building on a pre-trained text-to-image diffusion model, ControlVideo enhances the fidelity and temporal consistency by incorporating additional conditions(such as edge maps), and fine-tuning the key-frame and temporal attention on the source video-text pair via an in-depth exploration of the design space. Extensive experimental results demonstrate that ControlVideo outperforms various competitive baselines by delivering videos that exhibit high fidelity w.r.t. the source content, and temporal consistency, all while aligning with the text. By incorporating low-rank adaptation layers into the model before training, ControlVideo is further empowered to generate videos that align seamlessly with reference images. More importantly, ControlVideo can be readily extended to the more challenging task of long video editing(e.g., with hundreds of frames), where maintaining long-range temporal consistency is crucial. To achieve this, we propose to construct a fused ControlVideo by applying basic ControlVideo to overlapping short video segments and key frame videos and then merging them by pre-defined weight functions. Empirical results validate its capability to create videos across 140 frames, which is approximately 5.83 to 17.5 times more than what previous studies achieved. The code is available at https://***/thu-ml/controlvideo.

关键词： diffusion models controllable generation text-driven editing video editing long video editing

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：